creator-hub/internal/environment/store.go

package environment

import (
	"context"
	"crypto/rand"
	"database/sql"
	_ "embed"
	"encoding/hex"
	"encoding/json"
	"errors"
	"fmt"
	"net/url"
	"regexp"
	"sort"
	"strings"
	"time"
	"unicode/utf8"

	"github.com/jackc/pgx/v5/pgconn"
	_ "github.com/jackc/pgx/v5/stdlib"
)

//go:embed migrations/002_hub.sql
var migration002 string

//go:embed migrations/003_unified_accounts.sql
var migration003 string

//go:embed migrations/004_environment_actions.sql
var migration004 string

//go:embed migrations/005_sanitize_legacy_proxy.sql
var migration005 string

//go:embed migrations/006_runtime_cleanup.sql
var migration006 string

//go:embed migrations/007_runtime_binding_version.sql
var migration007 string

//go:embed migrations/008_runtime_cleanup_generation.sql
var migration008 string

//go:embed migrations/009_runtime_cleanup_compatibility.sql
var migration009 string

//go:embed migrations/010_runtime_network_generation.sql
var migration010 string

//go:embed migrations/011_task_recovery.sql
var migration011 string

//go:embed migrations/012_task_recovery_compatibility.sql
var migration012 string

//go:embed migrations/013_account_creation.sql
var migration013 string

//go:embed migrations/014_account_creation_compatibility.sql
var migration014 string

//go:embed migrations/015_gateway_rename_cascade.sql
var migration015 string

//go:embed migrations/016_network_exit_plain_credentials.sql
var migration016 string

//go:embed migrations/017_native_browser_versions.sql
var migration017 string

//go:embed migrations/033_unique_fingerprint_seed.sql
var migration033 string

//go:embed migrations/034_gateway_browser_default.sql
var migration034 string

//go:embed migrations/037_browser_env_image_version_repair.sql
var migration037 string

//go:embed migrations/001_phase_a.sql
var migration001 string

//go:embed migrations/038_creator_registration_backfill.sql
var migration038 string

//go:embed migrations/043_schema_consolidation.sql
var migration043 string

//go:embed migrations/1017_creator.sql
var migration1017 string

//go:embed migrations/1018_collection_checkpoints.sql
var migration1018 string

//go:embed migrations/1019_metric_plans.sql
var migration1019 string

//go:embed migrations/1020_competitor_sync_leases.sql
var migration1020 string

//go:embed migrations/1021_work_sources.sql
var migration1021 string

//go:embed migrations/1022_collection_lease_tokens.sql
var migration1022 string

//go:embed migrations/1023_password_references.sql
var migration1023 string

//go:embed migrations/1024_event_processing_times.sql
var migration1024 string

//go:embed migrations/1025_competitor_sync_tokens.sql
var migration1025 string

//go:embed migrations/1026_event_message_text.sql
var migration1026 string

//go:embed migrations/1027_creator_listener_state.sql
var migration1027 string

//go:embed migrations/1028_creator_event_message_type.sql
var migration1028 string

//go:embed migrations/1029_creator_source_sync_lease.sql
var migration1029 string

//go:embed migrations/1030_creator_event_gateway_time.sql
var migration1030 string

//go:embed migrations/1031_raw_payloads.sql
var migration1031 string

//go:embed migrations/1032_douyin_release_remediation.sql
var migration1032 string

//go:embed migrations/1033_competitor_tags.sql
var migration1033 string

//go:embed migrations/1034_competitor_tags_repair.sql
var migration1034 string

//go:embed migrations/1035_account_deletion.sql
var migration1035 string

//go:embed migrations/1036_competitor_unique_id.sql
var migration1036 string

//go:embed migrations/1037_competitor_share_jobs.sql
var migration1037 string

//go:embed migrations/1038_douyin_only_competitor_profile.sql
var migration1038 string

//go:embed migrations/1039_work_collect_count.sql
var migration1039 string

//go:embed migrations/1040_work_play_count.sql
var migration1040 string

//go:embed migrations/1041_creator_account_metric.sql
var migration1041 string

//go:embed migrations/1042_work_cover_cache.sql
var migration1042 string

var (
	ErrConflict          = errors.New("resource conflicts with existing state")
	ErrInvalid           = errors.New("invalid environment input")
	ErrNotFound          = errors.New("resource not found")
	ErrReconcileRequired = errors.New("runtime cleanup generation is unknown; manual reconciliation required")
)

func ValidNetworkExitID(id string) bool { return exitIDPattern.MatchString(id) }

var (
	aliasPattern       = regexp.MustCompile(`^[a-z0-9][a-z0-9-]{0,31}$`)
	gatewayNamePattern = regexp.MustCompile(`^[A-Za-z0-9][A-Za-z0-9._-]{0,63}$`)
	tokenPattern       = regexp.MustCompile(`^[A-Za-z0-9][A-Za-z0-9._-]{15,127}$`)
)

type Store struct {
	db            *sql.DB
	lockAdmission chan struct{}
}

// Gateway 是平台注册的 native browser gateway 节点；Token 由平台生成，明文存储供页面复制（开发阶段约定）。
type Gateway struct {
	Name      string    `json:"name"`
	Endpoint  string    `json:"endpoint"`
	Token     string    `json:"token"`
	CreatedAt time.Time `json:"created_at"`
	UpdatedAt time.Time `json:"updated_at"`
}

// Env 是一个浏览器环境；浏览器安装和默认运行时由 gateway 宿主机配置。
type Env struct {
	Alias          string      `json:"alias"`
	Name           string      `json:"name"`
	Gateway        string      `json:"gateway"`
	BrowserVersion string      `json:"-"` // legacy database test snapshot; runtime selection is gateway-owned
	Fingerprint    Fingerprint `json:"fingerprint"`
	CreatedAt      time.Time   `json:"created_at"`
}

func Open(ctx context.Context, databaseURL string) (*Store, error) {
	db, err := sql.Open("pgx", databaseURL)
	if err != nil {
		return nil, errors.New("open environment database")
	}
	db.SetMaxOpenConns(10)
	db.SetMaxIdleConns(2)
	db.SetConnMaxIdleTime(5 * time.Minute)
	if err := db.PingContext(ctx); err != nil {
		db.Close()
		return nil, errors.New("connect to environment database")
	}
	// LockResources keeps one connection until the lifecycle operation finishes.
	// Admit at most half the pool so those operations can still open nested DB calls.
	store := &Store{db: db, lockAdmission: make(chan struct{}, 5)}
	if err := store.migrate(ctx); err != nil {
		db.Close()
		return nil, err
	}
	return store, nil
}

func (s *Store) Close() error { return s.db.Close() }

func (s *Store) Ping(ctx context.Context) error { return s.db.PingContext(ctx) }

// LockResources serializes lifecycle state across control-plane replicas. The
// transaction carries no data changes; rolling it back only releases the locks.
func (s *Store) LockResources(ctx context.Context, aliases, exitIDs []string, _ ...[]string) (func(), error) {
	for _, alias := range aliases {
		if !aliasPattern.MatchString(alias) {
			return nil, ErrInvalid
		}
	}
	for _, id := range exitIDs {
		if !exitIDPattern.MatchString(id) {
			return nil, ErrInvalid
		}
	}
	if len(aliases)+len(exitIDs) == 0 {
		return func() {}, nil
	}
	select {
	case s.lockAdmission <- struct{}{}:
	case <-ctx.Done():
		return nil, ctx.Err()
	}
	tx, err := s.db.BeginTx(ctx, nil)
	if err != nil {
		<-s.lockAdmission
		return nil, errors.New("begin resource lock")
	}
	resources := []struct {
		namespace int
		keys      []string
	}{{1542738013, aliases}, {1542738015, exitIDs}}
	for _, resource := range resources {
		keys := append([]string(nil), resource.keys...)
		sort.Strings(keys)
		for _, key := range keys {
			if _, err := tx.ExecContext(ctx, `SELECT pg_advisory_xact_lock($1, hashtext($2))`, resource.namespace, key); err != nil {
				_ = tx.Rollback()
				<-s.lockAdmission
				return nil, errors.New("lock resource")
			}
		}
	}
	return func() {
		_ = tx.Rollback()
		<-s.lockAdmission
	}, nil
}

func (s *Store) migrate(ctx context.Context) error {
	tx, err := s.db.BeginTx(ctx, nil)
	if err != nil {
		return errors.New("begin environment schema migration")
	}
	defer tx.Rollback()
	if _, err := tx.ExecContext(ctx, `SELECT pg_advisory_xact_lock(1542738012)`); err != nil {
		return errors.New("lock environment schema migration")
	}
	if _, err := tx.ExecContext(ctx, `CREATE TABLE IF NOT EXISTS schema_migration (version integer PRIMARY KEY, applied_at timestamptz NOT NULL DEFAULT now())`); err != nil {
		return errors.New("create schema migration table")
	}
	for _, migration := range []struct {
		version int
		sql     string
	}{{1, migration001}, {2, migration002}, {3, migration003}, {4, migration004}, {5, migration005}, {6, migration006}, {7, migration007},
		{8, migration008}, {9, migration009}, {10, migration010}, {11, migration011}, {12, migration012}, {13, migration013}, {14, migration014},
		{15, migration015}, {16, migration016}, {17, migration017}, {33, migration033}, {34, migration034}, {37, migration037},
		{38, migration038},
		{1017, migration1017}, {1018, migration1018}, {1019, migration1019}, {1020, migration1020}, {1021, migration1021}, {1022, migration1022},
		{1023, migration1023}, {1024, migration1024}, {1025, migration1025}, {1026, migration1026}, {1027, migration1027}, {1028, migration1028},
		{1029, migration1029}, {1030, migration1030}, {1031, migration1031}, {1032, migration1032}, {1033, migration1033}, {1034, migration1034},
		{1035, migration1035}, {1036, migration1036}, {1037, migration1037}, {1038, migration1038}, {1039, migration1039}, {1040, migration1040},
		{1041, migration1041}, {1042, migration1042},
		{43, migration043}} {
		var applied bool
		if err := tx.QueryRowContext(ctx, `SELECT EXISTS (SELECT 1 FROM schema_migration WHERE version = $1)`, migration.version).Scan(&applied); err != nil {
			return errors.New("read environment schema migration state")
		}
		if applied {
			continue
		}
		if _, err := tx.ExecContext(ctx, migration.sql); err != nil {
			return fmt.Errorf("apply environment schema migration %d: %w", migration.version, err)
		}
		if _, err := tx.ExecContext(ctx, `INSERT INTO schema_migration (version) VALUES ($1)`, migration.version); err != nil {
			return fmt.Errorf("record environment schema migration %d", migration.version)
		}
	}
	if err := tx.Commit(); err != nil {
		return errors.New("commit environment schema migration")
	}
	return nil
}

func (s *Store) CreateGateway(ctx context.Context, name, endpoint, token string) (Gateway, error) {
	name, endpoint, token = strings.TrimSpace(name), strings.TrimSpace(endpoint), strings.TrimSpace(token)
	if !gatewayNamePattern.MatchString(name) || !validHTTPURL(endpoint) {
		return Gateway{}, ErrInvalid
	}
	// 未显式提供 token 时由平台生成；提供则与网关进程的 GATEWAY_TOKEN 保持一致。
	if token != "" {
		if !tokenPattern.MatchString(token) {
			return Gateway{}, ErrInvalid
		}
	} else {
		token = newToken()
	}
	gateway := Gateway{Name: name, Endpoint: endpoint, Token: token}
	err := s.db.QueryRowContext(ctx, `
		INSERT INTO gateway (name, endpoint, token) VALUES ($1, $2, $3)
		RETURNING created_at, updated_at`, name, endpoint, gateway.Token).
		Scan(&gateway.CreatedAt, &gateway.UpdatedAt)
	if err != nil {
		return Gateway{}, publicDatabaseError(err)
	}
	return gateway, nil
}

// UpdateGateway 修改网关名称、Endpoint 和令牌。名称变更由数据库外键 ON UPDATE CASCADE
// 原子同步 browser_env 引用；空令牌表示保留当前令牌，避免只改地址时意外轮换凭证。
func (s *Store) UpdateGateway(ctx context.Context, currentName, name, endpoint, token string) (Gateway, error) {
	currentName, name = strings.TrimSpace(currentName), strings.TrimSpace(name)
	endpoint, token = strings.TrimSpace(endpoint), strings.TrimSpace(token)
	if !gatewayNamePattern.MatchString(currentName) || !gatewayNamePattern.MatchString(name) || !validHTTPURL(endpoint) ||
		(token != "" && !tokenPattern.MatchString(token)) {
		return Gateway{}, ErrInvalid
	}
	var gateway Gateway
	err := s.db.QueryRowContext(ctx, `
		UPDATE gateway SET name = $1, endpoint = $2,
			token = CASE WHEN $3 = '' THEN token ELSE $3 END, updated_at = now()
		WHERE name = $4
		RETURNING name, endpoint, token, created_at, updated_at`, name, endpoint, token, currentName).
		Scan(&gateway.Name, &gateway.Endpoint, &gateway.Token, &gateway.CreatedAt, &gateway.UpdatedAt)
	if err != nil {
		return Gateway{}, rowError(err)
	}
	return gateway, nil
}

func (s *Store) ListGateways(ctx context.Context) ([]Gateway, error) {
	rows, err := s.db.QueryContext(ctx, `
		SELECT name, endpoint, token, created_at, updated_at FROM gateway ORDER BY created_at, name`)
	if err != nil {
		return nil, errors.New("read gateways")
	}
	defer rows.Close()
	gateways := []Gateway{}
	for rows.Next() {
		var gateway Gateway
		if err := rows.Scan(&gateway.Name, &gateway.Endpoint, &gateway.Token, &gateway.CreatedAt, &gateway.UpdatedAt); err != nil {
			return nil, errors.New("decode gateway")
		}
		gateways = append(gateways, gateway)
	}
	return gateways, rows.Err()
}

func (s *Store) GetGateway(ctx context.Context, name string) (Gateway, error) {
	var gateway Gateway
	if !gatewayNamePattern.MatchString(name) {
		return gateway, ErrInvalid
	}
	err := s.db.QueryRowContext(ctx, `
		SELECT name, endpoint, token, created_at, updated_at FROM gateway WHERE name = $1`, name).
		Scan(&gateway.Name, &gateway.Endpoint, &gateway.Token, &gateway.CreatedAt, &gateway.UpdatedAt)
	if err != nil {
		return Gateway{}, rowError(err)
	}
	return gateway, nil
}

func (s *Store) DeleteGateway(ctx context.Context, name string) error {
	if !gatewayNamePattern.MatchString(name) {
		return ErrInvalid
	}
	var deleted string
	if err := s.db.QueryRowContext(ctx, `DELETE FROM gateway WHERE name = $1 RETURNING name`, name).Scan(&deleted); err != nil {
		return rowError(err)
	}
	return nil
}

func (s *Store) CreateEnv(ctx context.Context, env Env) error {
	env.Alias = strings.TrimSpace(env.Alias)
	env.Name = strings.TrimSpace(env.Name)
	if !aliasPattern.MatchString(env.Alias) || !validDisplayName(env.Name) ||
		!gatewayNamePattern.MatchString(env.Gateway) || env.Fingerprint.ProxyServer != "" {
		return ErrInvalid
	}
	if err := env.Fingerprint.Validate(); err != nil {
		return fmt.Errorf("%w: %s", ErrInvalid, err)
	}
	encoded, err := json.Marshal(env.Fingerprint)
	if err != nil {
		return ErrInvalid
	}
	var created string
	err = s.db.QueryRowContext(ctx, `
		INSERT INTO browser_env (alias, name, gateway_name, fingerprint)
		VALUES ($1, $2, $3, $4)
		RETURNING alias`, env.Alias, env.Name, env.Gateway, encoded).Scan(&created)
	return rowError(err)
}

func (s *Store) ListEnvs(ctx context.Context) ([]Env, error) {
	rows, err := s.db.QueryContext(ctx, `
		SELECT alias, name, gateway_name, fingerprint, created_at
		FROM browser_env ORDER BY created_at, alias`)
	if err != nil {
		return nil, errors.New("read browser envs")
	}
	defer rows.Close()
	envs := []Env{}
	for rows.Next() {
		env, err := scanEnv(rows)
		if err != nil {
			return nil, err
		}
		envs = append(envs, env)
	}
	return envs, rows.Err()
}

func (s *Store) GetEnv(ctx context.Context, alias string) (Env, error) {
	if !aliasPattern.MatchString(alias) {
		return Env{}, ErrInvalid
	}
	rows, err := s.db.QueryContext(ctx, `
		SELECT alias, name, gateway_name, fingerprint, created_at
		FROM browser_env WHERE alias = $1`, alias)
	if err != nil {
		return Env{}, errors.New("read browser env")
	}
	defer rows.Close()
	if !rows.Next() {
		return Env{}, ErrNotFound
	}
	env, err := scanEnv(rows)
	if err != nil {
		return Env{}, err
	}
	return env, rows.Err()
}

func (s *Store) DeleteEnv(ctx context.Context, alias string) error {
	if !aliasPattern.MatchString(alias) {
		return ErrInvalid
	}
	var deleted string
	if err := s.db.QueryRowContext(ctx, `DELETE FROM browser_env WHERE alias = $1 RETURNING alias`, alias).Scan(&deleted); err != nil {
		return rowError(err)
	}
	return nil
}

// DeleteAccountEnvironment removes the account binding and its private browser environment.
// Shared gateways, images, and network exits are intentionally preserved.
func (s *Store) DeleteAccountEnvironment(ctx context.Context, accountID string) error {
	if strings.TrimSpace(accountID) == "" {
		return ErrInvalid
	}
	tx, err := s.db.BeginTx(ctx, nil)
	if err != nil {
		return errors.New("begin account environment deletion")
	}
	defer tx.Rollback()
	var alias string
	if err := tx.QueryRowContext(ctx, `
		SELECT alias
		FROM browser_env
		WHERE account_id = $1
		FOR UPDATE`, accountID).Scan(&alias); errors.Is(err, sql.ErrNoRows) {
		return nil
	} else if err != nil {
		return errors.New("read account environment binding")
	}
	if _, err := tx.ExecContext(ctx, `DELETE FROM browser_env WHERE alias = $1`, alias); err != nil {
		return errors.New("delete account browser environment")
	}
	return commitHub(tx)
}

func scanEnv(rows *sql.Rows) (Env, error) {
	var env Env
	var encoded []byte
	if err := rows.Scan(&env.Alias, &env.Name, &env.Gateway, &encoded, &env.CreatedAt); err != nil {
		return Env{}, errors.New("decode browser env")
	}
	if len(encoded) > 0 {
		if err := json.Unmarshal(encoded, &env.Fingerprint); err != nil {
			return Env{}, errors.New("decode env fingerprint")
		}
	}
	env.Fingerprint.ProxyServer = ""
	env.Fingerprint.DisableNonProxiedUDP = false
	return env, nil
}

func validDisplayName(name string) bool {
	if name == "" || utf8.RuneCountInString(name) > 64 {
		return false
	}
	for _, r := range name {
		if r < 0x20 || r == 0x7f {
			return false
		}
	}
	return true
}

func validHTTPURL(raw string) bool {
	parsed, err := url.Parse(raw)
	return err == nil && parsed.Host != "" && (parsed.Scheme == "http" || parsed.Scheme == "https")
}

func newToken() string {
	var value [24]byte
	_, _ = rand.Read(value[:])
	return hex.EncodeToString(value[:])
}

func commitHub(tx *sql.Tx) error {
	if err := tx.Commit(); err != nil {
		return errors.New("commit environment transaction")
	}
	return nil
}

func rowError(err error) error {
	if errors.Is(err, sql.ErrNoRows) {
		return ErrNotFound
	}
	return publicDatabaseError(err)
}

func publicDatabaseError(err error) error {
	if err == nil {
		return nil
	}
	var postgresError *pgconn.PgError
	if errors.As(err, &postgresError) && (postgresError.Code == "23505" || postgresError.Code == "23503" || postgresError.Code == "23514") {
		return ErrConflict
	}
	return errors.New("environment persistence operation failed")
}